Overmind raises £2m seed to secure autonomous AI agents with reinforcement learning
This article covers Overmind, an AI startup that has raised £2m in a seed funding round led by Osney Capital to develop tools that observe, secure and improve agent behaviour as organisations deploy autonomous models in real environments. The funding is intended to support development of operational monitoring and control for agentic AI, targeting regulated and high-risk sectors such as legal, healthcare and fintech.
Overmind, an AI startup building a supervision layer for agentic AI systems, has raised £2 million in a seed funding round led by Osney Capital to develop tools that observe, secure and improve agent behaviour as organisations deploy autonomous models in real environments.
Why it matters
As more companies move from experimental AI demos to systems that act autonomously, existing security and monitoring tools struggle to keep up. Gartner warns that 40% of agentic AI projects will be cancelled by 2027 largely because of inadequate risk controls, a signal that operational failures rather than model accuracy may determine adoption. Overmind positions itself at that operational layer, aiming to give teams visibility and real-time control over agents before errors cascade.
Inside the product
Overmind describes its platform as a supervision layer that watches agent behaviour, detects deviations and intervenes in real time. The system uses reinforcement learning to not only block unsafe actions but also iteratively improve agent performance and accuracy over time. The company plans to target regulated or high-risk verticals where autonomous actions can have material consequences, including legal, healthcare and fintech.
Who are the investors?
The round was led by Osney Capital, with participation from 14Peaks, Portfolio Ventures, Antler and Endurance Ventures.
In the announcement, Adam Cragg, Partner at Osney Capital, said:
In the new frontier of autonomous AI, agent security, performance, and execution are the ultimate competitive advantages. Overmind provides businesses with truly differentiated technology that monitors and secures agentic AI while iteratively improving model performance, enabling teams to scale with confidence. We’re excited to back such a strong founding team addressing a critical market.
Osney is the lead backer and other participants are a mix of early-stage and venture investors. Their involvement signals continued interest in infrastructure that addresses operational risk as firms adopt agentic AI.
If you're researching potential backers in this space:
- Explore our list of AI venture capital firms in the UK
- Browse our directory of AI angel investors in the UK
Founder perspective
The company was founded by Tyler Edwards and Akhat Rakishev. Edwards spent eight years building AI systems for British intelligence agencies; Rakishev previously led machine learning infrastructure teams at Monzo and Lyst. That combination of public-sector operational experience and fintech production engineering shapes Overmind’s focus on safety, observability and reliability.
In the announcement, Tyler Edwards, Co-founder & CEO at Overmind, said:
The AI security industry is trying to secure the wrong thing. Model security matters, but the deeper challenge is operational. AI is already acting at a speed and scale humans never could. We need infrastructure that can keep pace, not just catching failures, but continuously learning from how agents behave in the real world. That's what makes the difference between impressive demos and technology you can trust.
The bigger picture
This raise arrives as investors and buyers rethink priorities for AI deployment. Tools that manage execution risk—monitoring what agents do and automatically correcting behaviour—are emerging as a distinct category from model-centric security. For startups developing autonomous systems, proving robust operational controls will likely be a prerequisite for regulatory approval and enterprise adoption.
The deal also reflects broader appetite among AI investors for infrastructure businesses that enable safe, scalable use of advanced models. As agentic systems proliferate across industries, companies that can bridge the gap between model capability and operational reliability may find steady demand.
Overmind’s seed round is another example of early-stage funding flows in the UK AI scene, where teams with production engineering and public-sector security experience are drawing capital to build the tooling enterprises will need as autonomous systems move into everyday workflows across Europe.
They've invested in Overmind
Get to know these AI investors
Click here for a full list of 7,526+ startup investors in the UK